Writing study: POLAR: A Portrait OLAT Dataset and Generative Framework for Illumination-Aware Face Modeling
← CVPR 2026 oral writing corpus
Zhuo Chen, Chengqun Yang, Zhuo Su, Zheng Lv, Jingnan Gao, Xiaoyuan Zhang, Xiaokang Yang, Yichao Yan · Official CVPR 2026 oral record
Figure, table and section ledger
Editorial contribution type: Datasets, benchmarks & measurement (primary).
Layout extraction, not a required paper template. Counts are distinct caption labels; a multi-panel figure counts once. Section spans are estimated from heading positions, including floats and whitespace. Method and limitations.
- Canonical PDF
- 11 physical pages
- Before references / appendix
- ≈ 8.1 page-equivalents
- Main figures / tables
- 9 / 2
- Attached appendix figures / tables
- 0 / 0
- Attached appendix span
- Separate file recorded below; not in this PDF total
Extraction notes: No attached appendix boundary detected. Zero means no recognized caption, not independently verified absence.
Section-by-section lengths and source pages
The title/abstract/front matter precede the first numbered heading. A span ends at the next detected top-level heading, not at the last sentence; end-page whitespace is included. References and statements are listed separately. These are not measurements of text-only or artwork area.
| Top-level heading | Starts on PDF page | Estimated span | Share of entire PDF |
|---|---|---|---|
| 1. Introduction | 1 | ≈ 1.01 pp. | 9.18% |
| 2. Related Work | 2 | ≈ 0.51 pp. | 4.64% |
| 3. The POLAR Dataset | 3 | ≈ 1.89 pp. | 17.18% |
| 4. Flow-based OLAT Generative Model | 5 | ≈ 1.92 pp. | 17.45% |
| 5. Experiments | 7 | ≈ 1.58 pp. | 14.36% |
| 6. Conclusion and Discussion | 8 | ≈ 0.29 pp. | 2.64% |
| References | 9 | ≈ 2.9 pp. | 26.36% |
All detected figure and table locations
Main text
Attached appendix
No recognized caption labels.
Separate source files—excluded from canonical-PDF totals
A separate author version may repeat main text and references. Whole-file lengths and counts are not extra appendix material and are never added to the canonical PDF totals. Read the study's source scope for which sections received substantive reading versus layout inspection.
Additional source 1
Separate PDF (or ZIP containing it): whole-file layout of 8 physical pages; 13 detected figures and 1 tables. These may include overlapping main-text material.
Located appendix only: starts on physical p.1; ≈ 7.93 page-equivalents, 13 figure labels and 1 table labels. This interval is distinct from the whole file.
Argument and organization
- Capture-to-generation loop
Describe the physical dataset, processing and comparison before the learned lighting transport model.
- Generated versus captured evidence
Separate generated OLAT responses, composed relighting and downstream baseline comparisons.
Problem, claim and contribution
The data/model loop is explicit: real calibrated captures train a generator that expands subject diversity, rather than treating synthetic samples as equivalent original observations.
Introduction and Figure 2 · Main paper · Additional source 2
Argument and section order
Acquisition, modalities and processing have their own subsections before bridge matching and experiments, making data provenance part of the argument.
Sections 3.1–4.2 · Main paper · Additional source 2
Figures, captions and visual explanation
Figure 4 presents controlled visual processing variants; Figure 6 groups predicted per-light responses separately from relit outputs and other methods.
Figure 4, PDF p. 4; Figure 6, p. 7 · Main paper · Additional source 2
Tables, comparisons and evidence
Table 2 retains different winners across perceptual, PSNR and SSIM metrics; the experimental text distinguishes composition from captured OLATs versus generated OLATs.
Table 2 and Section 5.1, PDF p. 7 · Main paper · Additional source 2
Appendix and reproducibility
The supplementary study distinguishes real captures from predicted OLATs, adds delighting/objective controls, and reserves implementation details as a separate section.
Supplement §§A–D; Table 1 · Main paper · Additional source 2
Clarity and writing conventions
The paper repeatedly labels captured, generated and composed quantities, helping readers track which evidence measures physical capture fidelity versus generative plausibility.
Figure 7 and Section 5.1 · Main paper · Additional source 2
What to reuse—and what not to infer
Transferable lesson: For a dataset-plus-generator paper, preserve a visible provenance chain from capture to synthesis and keep photometric fidelity, perceptual quality and coverage distinct.
Boundary: An in-the-wild comparison can use a reference under the same lighting without sharing identity; captions must make reference meaning explicit. Not every reported metric favors the proposed method.
Apply this lesson using the argument and evidence worksheet.
Inspection record and sources
- Checked
- 2026-09-26
- Reading status
- Writing-focused manuscript scan + selected visuals inspected
- Main text
- Manuscript-wide writing scan of main PDF pp. 1–8: framing, section structure, method exposition, figure/table captions, evaluation organization and conclusions; not an exhaustive technical proof audit or reproduction.
- Appendix / supplement
- Scanned the eight-page official ZIP-member PDF for acquisition, expression/statistical coverage, HDR examples, processing, delighting and objective ablations, real/predicted OLAT comparisons, relighting consistency, and implementation details; video/code unexecuted.
- Visually inspected PDF pages
- 4, 7
This is a writing-focused scan of the manuscript and the stated supplementary scope: framing, organization, visual explanation, evidence presentation and wording. It is not a line-by-line proof audit, reproduction of results or exhaustive scientific review.